An Observational, Retrospective, Cohort Study of Dosing Patterns for Rofecoxib and Celecoxib in the Treatment of Arthritis

Objective: This study assessed prescribing patterns for rofecoxib and celecoxib in the treatment of osteoarthritis (OA) and rheumatoid arthritis (RA), as well as differences in prescribing patterns across physician specialties. Methods: This was an observational, retrospective, cohort study of a large, US pharmacy claims database. Eligible patients were initiating therapy with rofecoxib or celecoxib and had ≥90 days' supply of medication, as well as ≥1 medical claim specific to OA or RA between June 1, 2000, and May 31, 2001. Analyses were stratified according to diagnosis, prescribing physician specialty, and patient demographics. The main outcome measure was mean daily usage (ie, mean daily dose [milligrams]; mean number of pills per day; and mean daily consumption [denoted as DACON], calculated as daily dose divided by most frequently prescribed strength). This was primarily a descriptive study. Tests of statistical significance were not performed because the large sample size would have rendered small differences significant. Results: A total of 58,574 patients with OA (81.8% [n = 47,935]) or RA (18.2% [n = 10,639]) received 220,627 prescriptions for rofecoxib or celecoxib (47.7% [n = 27,924] and 52.3% [n = 30,650] of patients, respectively) during the study period. Overall, the most frequently prescribed strengths were rofecoxib 25 mg and celecoxib 200 mg. In both OA and RA, the most frequently prescribed mean daily dose of rofecoxib was 25 mg. In OA, the most frequently prescribed mean daily dose of celecoxib was 200 mg; in RA, it was 400 mg. Both pills per day and DACON were higher for celecoxib than rofecoxib. The DACON for rofecoxib was unrelated to physician specialty. Rheumatologists prescribed celecoxib at 20% to 40% higher mean daily doses than did primary care physicians, orthopedic specialists, or other specialists. Regardless of physician specialty, the DACON appeared higher for patients with RA than OA, for men than women, and for younger (aged <65 years) than older patients. Conclusions: In this analysis, relative to the most frequently prescribed strength, celecoxib-treated patients with OA and RA had higher DACONs than rofecoxib-treated OA and RA patients across all subgroups. These observations may have economic implications in terms of direct effects on cost and the need for formularies to consider overall use patterns in addition to pill costs. However, these conclusions are limited by lack of clinical information (other than an OA or RA diagnosis), inability to ascertain actual use, and potential for selection bias.